In Vitro Activity of Sanfetrinem and Affinity for the Penicillin-Binding Proteins of Streptococcus pneumoniae

Abstract

Against penicillin-susceptible pneumococci, the activity of sanfetrinem was similar to those of penicillin, amoxicillin, cefotaxime, imipenem, and meropenem, while against penicillin-resistant strains, sanfetrinem and the carbapenems exhibited superior activity (MICs at which 90% of strains are inhibited, ≤1 μg/ml). PBP 1a in the penicillin-susceptible strain and PBP 1a and PBP 2b in the more resistant isolates seemed to be the essential penicillin-binding proteins for imipenem and sanfetrinem.

The increasing number of penicillin-resistant strains and their resistance to multiple antibiotics complicates the strategy for the treatment of pneumococcal infections such as otitis media (4). Thus, compounds with bactericidal activity superior to those in current use are needed. Among the β-lactam antibiotics, compounds which have comparable MICs may differ greatly in their bactericidal potencies: penicillin and carbapenem can rapidly kill and lyse pneumococcal strains, while cephalosporins have only poor lytic activity (5, 6).

In this study, we determined the in vitro susceptibilities of penicillin-susceptible, intermediately penicillin-resistant, and penicillin-resistant strains of Streptococcus pneumoniae to the first member of a new class of oral tricyclic β-lactam compounds (trinems), sanfetrinem (formerly GV104326) (3, 11, 12). We also studied its affinity for the penicillin-binding proteins (PBPs) of selected strains and its bactericidal and lytic activities.

The bacteriostatic effect of sanfetrinem against 70 clinical isolates with different levels of resistance to penicillin was compared with those of other β-lactam antibiotics (Table 1). MICs were determined on Mueller-Hinton agar containing 5% horse blood, with a Steers replicator device delivering 10⁴ to 10⁵ CFU per spot. The MICs were read after 18 h at 37°C. Against the penicillin-susceptible strains, the activity of sanfetrinem was similar to those of penicillins, cefotaxime, meropenem, and imipenem. Against penicillin-resistant strains, the carbapenems and sanfetrinem exhibited superior activity, particularly against highly penicillin-resistant strains, with MICs at which 90% of strains are inhibited (MIC₉₀s) of ≤1 μg/ml. Sanfetrinem was in general twofold less active than imipenem against the resistant strains, with its activity being intermediate between those of imipenem and meropenem. These results are in agreement with those recently reported (11, 12).

TABLE 1.

In vitro activities of different β-lactam antibiotics against 70 clinical S. pneumoniae isolates

Antibiotic	MIC (μg/ml)a
	PSP (n = 19)			IPRP (n = 21)			PRP (n = 30)
	50%	90%	Range	50%	90%	Range	50%	90%	Range
Penicillin	0.015	0.015	0.007–0.06	0.25	1	0.125–1	2	2	2–4
Amoxicillin	0.015	0.015	0.007–0.015	0.125	0.25	0.03–2	2	2	1–4
Cefotaxime	0.015	0.03	0.007–0.03	0.125	0.25	0.06–0.5	1	2	0.5–4
Meropenem	0.007	0.03	0.007–0.06	0.125	0.25	0.03–0.5	0.5	1	0.5–1
Imipenem	0.007	0.015	≤0.003–0.015	0.03	0.125	0.015–0.25	0.25	0.25	0.125–0.25
Sanfetrinem	0.007	0.015	≤0.003–0.03	0.06	0.25	0.03–0.5	0.5	0.5	0.25–1

^aPSP, penicillin-susceptible pneumococci; IPRP, intermediately penicillin-resistant pneumococci; PRP, penicillin-resistant pneumococci. 

The lytic and bactericidal activities of sanfetrinem were studied against three selected strains. S. pneumoniae R6 is a susceptible derivative of the unencapsulated Rockefeller University strain R36A; S. pneumoniae 4387 and S. pneumoniae 4411 are intermediately penicillin-resistant and penicillin-resistant clinical isolates, respectively (10); both strains belong to serotype 23F. The strains were grown at 37°C in Todd-Hewitt broth supplemented with 0.5% (final concentration) yeast extract (Difco Laboratories, Detroit, Mich). Growth and lysis were monitored by measuring the optical density at 620 nm (OD₆₂₀). When an absorbance of 0.30 to 0.35 was reached (10⁷ to 10⁸ CFU/ml), penicillin G, imipenem, or sanfetrinem was added at a concentration of four times the MIC. Viable bacteria were counted 2 and 4 h after antibiotic addition by plating appropriately diluted cultures into Columbia agar containing 5% horse blood.

The mode of action of penicillin has been extensively studied (13). Moreillon et al. (8) showed that the bactericidal effect of penicillin can result from the triggering of two independent systems: the autolytic system (Lyt⁺) which involves an amidase, the major pneumococcal autolysin, and an autolysin-independent system (Cid⁺) involving cid determinant, the role of which is not known. Significant lysis was obtained at four times the MIC of each antibiotic (Fig. 1) for penicillin-susceptible strain R6 and intermediately penicillin-resistant strain 4387, but the lysis was more pronounced for the latter strain. A significant bactericidal effect was associated with lysis, which resulted in a decrease of at least 4 log CFU after 4 h and which again was more pronounced for strain 4387. Interestingly, the three compounds showed only slight lytic effects against penicillin-resistant strain 4411, while their bactericidal effects were still significant and nearly reached that obtained for R6. Similar results were obtained with each antibiotic at eight times the MIC, while at twice the MIC, each compound was slightly less effective (data not shown). According to the classification of Moreillon and colleagues (8, 9), these results suggest that, as for strain R6, the intermediately penicillin-resistant strain 4387, which is lysed by β-lactams, is of the Tol⁻ (nontolerant) Lyt⁺ Cid⁺ phenotype, while the penicillin-resistant strain 4411, which was lysed only in the presence of Triton X-100 due to the presence of amidase (data not shown), appears to be tolerant (Tol⁺) and therefore to be of the Tol⁺ Lyt⁺ Cid⁺ phenotype. It was previously shown that among the β-lactam antibiotics, penems and carbapenems have superior lytic and bactericidal activities against S. pneumoniae (1, 2, 6, 12), even against tolerant strains (Tol⁺ Lyt⁻ Cid⁺ and Tol⁺ Lyt⁺ Cid⁻) (1, 6, 7). This also seems to be the case for the trinem sanfetrinem.

FIG. 1.

Lysis curves and killing curves showing the effects of sanfetrinem (SAN), imipenem (IMI), and penicillin G (PEN) against the susceptible pneumococcal reference strain R6 (□; MICs: penicillin, 0.015 μg/ml; imipenem, 0.003 μg/ml; sanfetrinem, 0.003 μg/ml), intermediately resistant strain 4387 (○; MICs: penicillin, 0.125 μg/ml; imipenem, 0.03 sanfetrinem, 0.03 μg/ml), and penicillin-resistant strain 4411 (▵; MICs: penicillin, 2 μg/ml; imipenem, 0.25 μg/ml; sanfetrinem, 0.5 μg/ml). Antibiotics at four times the MIC were added at time zero. The control culture (⊡) received no antibiotic.

Analysis of PBPs was done exactly as described previously (10) by using [³H]benzylpenicillin (26 Ci/mmol), which was a generous gift from Rhône-Poulenc-Rorer Research (Vitry sur Seine, France). These conditions allowed for the optimal separation of three PBPs, PBPs 2x, 2a, and 2b. To obtain a better separation of PBP 1a and PBP 1b, an acrylamide/bisacrylamide ratio of 30:0.8 with a 10% separating gel was used. Competition experiments with increasing concentrations of imipenem or sanfetrinem showed that in susceptible strain R6, 50% saturation of PBP 1a and PBP 3 was obtained close to the MICs of these antibiotics (Table 2), while in intermediately resistant strain 4387 and resistant strain 4411, 50% saturation of PBP 1a and PBP 2b was observed under these conditions. Therefore, PBP 1a in penicillin-susceptible strain R6 and both PBP 1a and PBP 2b in the more resistant isolates are the most likely essential PBPs for imipenem and sanfetrinem. Finally, the preferential target for sanfetrinem, as for many β-lactam antibiotics (14), appears to be PBP 1a. Since PBP 2b has been involved as a major target for triggering lysis in the presence of penicillin (15), this would explain why lysis was already observed at twice the MIC for the Tol⁻ Lyt⁺ Cid⁺ intermediately resistant strain 4387. In the case of the Tol⁺ Lyt⁺ Cid⁺ resistant strain 4411, the process which should trigger lysis is probably independent of the level of PBP saturation (8). Concerning susceptible strain R6, one might think that saturation of both PBP 1a and PBP 3 was sufficient to trigger lysis.

TABLE 2.

Fifty percent saturation of PBPs of susceptible strain S. pneumoniae R6, intermediately resistant strain S. pneumoniae 4387, and resistant strain S. pneumoniae 4411 expressed as a multiple of the MIC of imipenem and sanfetrinem

PBP	Multiple of the MICa
	R6		4387		4411
	IMI	SAN	IMI	SAN	IMI	SAN
1a	1	1	0.35	1.5	0.5	1.5
1b	6	3	—b	—	—	—
2x	6	6	4	6	8	6
2a	4	12	—	—	0.06	0.125
2b	6	6	0.75	3	1.5	0.5
3	1.5	1.5	0.125	0.16	0.06	0.03
MIC (μg/ml)	0.003	0.003	0.03	0.03	0.25	0.5

^aIMI, imipenem; SAN, sanfetrinem. 

^b—, PBP not visualized by direct labeling with ³H-benzylpenicillin. 

In conclusion, since sanfetrinem, a new oral trinem, showed a bactericidal effect at concentrations very close to its MIC, it may be of use for the treatment of pneumococcal infections, particularly otitis media, provided that it reaches adequate levels to eradicate intermediately and fully penicillin-resistant pneumococci.